Most recent experiments have indicated that distal pocket polarity rather than steric hindrance is the major factor governing the distribution of FeCO stretching frequencies (nu(C-O), nu(Fe-CO)) in myoglobins and hemoglobins. Hydrogen bonding and other polar interactions have also been shown to play a key role in regulating O-2 and CO binding. To quantify the effects of polarity on nu(C-O), nu(Fe-CO), and ligand binding, we calculated electrostatic potential field distributions in the distal pockets of 18 different mutants and two wild-type forms of recombinant pig and sperm whale MbCO. The results were obtained using linearized Poisson-Boltzmann methods with coordinates from high-resolution structures determined experimentally by X-ray crystallography. The computed potential fields at the ligand atoms vary from +30 to -12 kcal/mol depending on the protein structure at the distal site. The electrostatic fields correlate inversely with nu(C-O) and directly with nu(Fe-CO) In all our calculations, the distal histidine is modeled as the neutral N delta-H tautomer, regardless of which ferrous ligand is bound. If the neutral N epsilon-H tautomer is used, the computed potentials at the bound ligand atoms are uniformly negative and show no correlation with nu(C-O), nu(Fe-C), and any ligand binding parameters. Although calculated using primarily MbCO structures, there is a linear, inverse relationship between the electrostatic field at the ligand binding site and the logarithm of the rate constant for O-2 dissociation. As a result, high O-2 affinity can be predicted semiquantitatively from a large positive potential field or from an experimentally low value of nu(C-O). Thus, the stretching frequency of bound CO serves as an empirical voltmeter that can be used to measure the polarity of the distal pocket and to predict the extent of electrostatic stabilization of bound O-2.